Markov Switching Granger Causality

Theories on the relationship between money and inflation had largely been shaped around the positive relationship and money causality for inflation before the Post-Keynesians. Since the 1980s, this idea emerged that there might be no correlation between money growth and inflation. In the case of existence, the causality is reversed, so money is endogenous somehow. However, practically there is a suspicion that the causality between money growth and inflation is not fixed and linear. According to the experience of Turkey in the last seven decades, which has experienced fluctuated inflation rates, it is supposed that the causal relationship between Broad Money Growth (BMG) and inflation is not constant. This paper examines this idea over 1961–2019 using a Markov Switching Vector Autoregressive model (MS-VAR), which allows for regime shifts. Findings show that the causal relationship between BMG and inflation has not been constant, and different regimes have generated different causality orientations. There was a one-way causality from inflation to BMG during 1971–2001 that inflation rates were high. Whereas, during 1961–70 and 2002–19, when the Turkish economy experienced milder inflation rates, there was a one-way causal relationship from BMG to inflation.